Introduction:
In 2006, the Centers for Disease Control and Prevention (CDC) began directly estimating human immunodeficiency virus (HIV) incidence based upon the Serologic Testing Algorithm for Recent HIV Seroconversion (STARHS) and individuals’ testing and treatment history, collected as part of the HIV Incidence Surveillance (HIS) system. The algorithm relies largely upon individuals’ self-reported testing history. The primary objective of this study is to describe the methodology and procedures used to assess the similarities between self-reported and medical record data on HIV status and the dates of first positive and last negative tests.

Material and methods:
The testing history from the individuals is used in combination with the latest laboratory assay tests to obtain a direct population-based estimate of HIV incidence. Understanding how accurate the self-report testing information is will help in estimating HIV incidence. Partnerships were made with the medical clinics and Counselling and Testing facilities, and the participants were recruited from the patients attending these facilities. Participants were interviewed to ascertain the self-report of HIV test, HIV status, the date and location of the most recent HIV-negative test and the first HIV-positive test. Medical record abstraction was done after participants’ authorisation to compare the accuracy of self-report testing data.

Results:
The participants’ response rate was 83.7%, and the most common reason of not participating was lack of interest in the study. Data analysis is reported by CDC and Houston Health Department in a separate article.

Conclusions:
The methodology and procedures adopted in this study can be adopted, replicated, and improved in future studies exploring self-report accuracy.